Storage systems, such as distributed storage systems, ensure availability and reliability of data by using replication and redundant arrays of independent disks (RAID).
In such a distributed storage system, when a disk (storage device) becomes faulty, operations for recovery of the faulty disk are performed in the background of user applications.
Related techniques are disclosed in, for example, Japanese Laid-open Patent Publication No. 2006-259894, Japanese Laid-open Patent Publication No. 2008-186400, and Japanese National Publication of International Patent Application No. 2013-506191.
Recovery of a faulty disk involves data transfer from a normal disk. A network band is therefore used for recovery of the faulty disk. This affects the user applications in some cases. When a plurality of disks have become faulty, operations of recovering a faulty disk are performed independently for each faulty disk. For this reason, there are some cases where a larger network band is used for recovery and where a network band is used for recovery for a longer time.